Problem or bug: offsets on inputs?

i have some very strange tuning problems…
when having a custom unit osc on different tracks they are out of tune among themselves.

so why is that?
i found out that every input on the er301 has a static offset. a different one… at least my unit.

i have an input assigned to the freq of the oscillator for fm purpose.
Nothing is plugged in into that input.
if all oscillators have the same fm input, fine. of course.
but if one oscillator has a different fm input its out of tune to the rest.
so there must be some kind of offset to the inputs.
interestingly its not all of them. some share the same, and some are different. but those that are different are not all the same.
very strange. thats a huge problem though when you want to use it as an osc.
@odevices any idea whats wrong here?

Mine definitely doesn’t do that. Just to make sure, you’re using the inputs labeled ABCDx for V/O, right? Those are the ones calibrated for it.

1 Like

no, i´m using ABCDx for f0…
i want linear fm…

interestingly when i route different inputs to the V/O theres not offset… so this is a software or a user problems:slight_smile:

So if you have the capability:

Two sin units each inside a mixer on the same chain.

Send a fixed voltage through a buffered mult into two different inputs (ABCDx), and assign them each to f0.

Make sure offset and gain for f0 is set the same on each.

Do the oscillators go out of tune then?

Just throwing out some ideas to help isolate. Take FM out of the equation.

As far as I know there’s no guarantee that the sine osc will be in sync, so if you are using them to FM you will almost certainly develop inconsistencies.

Sync input will get them all cycling in the same phase and from there you will be able to modify meaningfully.

ok, thats a complicated problem…
phase sync is not an issue for me.

i want to have the (Edit:) IN1 or IN2 or IN3 or IN4 input assigned to the f0. and if i do this the osc get out of tune.
with nothing plugged into the in inputs.
but maybe i´m just too tired… will have a look into this tomorrow…

Okay - yeah sounds weird.

So two different inputs, does the behaviour reverse when you swap the inputs over?

Can carry on tomorrow, if you’re tired never a good time to try and debug something :slight_smile:

After reading through this thread, I was still left confused about which (external) inputs you were using. Can you describe the patch one more time but be more specific about the inputs please? :bowing_man:

The inputs on the ER-301 are called:

  • G1-G4: for gates
  • IN1-IN4: for audio and non-calibrated CV
  • A1-A3, B1-B3, C1-C3, and D1-D3: for calibrated CV

yes, sorry, the bug is very confusing and it makes me think of that i just do a simple error i oversee all the time… maybe i can load up a quicksave, is this easy?

my problem is, i have this problem with a simple patch i build. if i build something simpler again from scratch the bug will not happen.
but i had this bug in this patch and a patch i did some time ago.

so, let me try to describe what i did:
i have 4 custom units, 1 on every channel.
all the custom units are nearly the same. they consist of two mixer channels with two osc + vca in it. (sine and saw in one, sine and triangle in the other, triangle and saw in the third, and the fourth channel has the same as channel 1 ).
I have in the custom unit 3 controls. these custom controls are assigned to both oscillators in that cutsom unit.
custom unit v/oct controll is assigned to v/oct of the two oscillators, of course.
custom unit linear control is assigned to f0 of the two oscillators.
custom unit db control is assigned to the vca. (i morph this way between the two oscillators in the custom unit.)
(allthough this last part does not have anything to do with the problem…)

the v/oct in at the custom unit is A1 for the first channel, A2 for the second channel, A3 for the third, B1 for the fourth.
the linear control is assigned to the IN1 for channel 1, IN2 for channel 2 and so on up to channel 4.

those 4 channels are really devided by each other, no control is shared.

right now channel 1 and 2 share the same tuning. channel 4 is way out of tune.

ok, the bias is all the same for the mod amount. key to that it should be the same.
the (multiplication (X)) gain in the mod page is also the same in all modules.

now here comes the, for me, strange thing. if i change the miltiplication gain for the modulation for the f0, i tune the pitch.
please note, ALL the physical inputs are empty and NOT patched.
in my understand a multiplication gain does NOTHING when nothing is patched in…

next strange thing, and this can be easy replicated with every patch from scratch, when i change the input to the f0 from from IN1 to an A1-A3 or any other, i get a different pitch say a different offset.

When i work from scratch at least the IN1 - IN4 seem to have the same offset and the A1-A3, B1-B3, C1-C3, and D1-D3 share the same.
But as i wrote above, not everytime.

hope this was clear somehow.
and maybe i should just upload the quicksave.
(can i just grab this from the card?)

thanks a lot!

The first suspicious thing that I noticed right away is the use of the IN1-IN4 inputs for the very sensitive f0 (linear FM) parameter. The IN1-4 inputs are not calibrated and thus might have small offsets. This is due to the analog anti-aliasing filters that are installed in their signal path. You should first forget about using the IN1-IN4 inputs for precision CV.

I will also try to replicate your patch (using only the ABCD inputs) and report back my findings.

i want to use them to have a higher resolution for FM…

ok, i tried one more thing, and now i think i know whats the problme. maybe :slight_smile:
there are indeed very small different offset on every input.
very very small

i changed the input for the f0 to C1 C2 C3 and D1.
First it was relative in tune. i say relative cause it still had an unisono sound, so its slightly out of tune, but i could live with it.
changing the mod amount (mod gain multiplication) still changes the pitch, so there has to be an offset i thought.
and than i confirmed that to me when i raised the mod amount. it all got totally out of tune.
so whats the deal? when having the system as descibed, i have two mod amounts (gain moultiplications.)
one on the oscillator itself on the f0 input, and on on the custom unit for the f0.
when i raise both to +5 i get a lot of gain multiplication. and there you gain the offsets that are there on each of all inputs a lot. and than you hear that all inputs have a different offset…
but in practice, you never want two x5 mulitplications of the mod input.
so all in all this is more a user error. i don´t need that much, and having discovered that it just showed me a flaw of the unit we don´t have to be too much concerned about as it will most probably never come in our way. never.

but its good to know!
i´m happy to move on as nothing had happened :slight_smile:

FYI, the IN1-IN4 and ABCD inputs have the same resolution and sample rate.

1 Like

"Can be used to record hiqh-quality audio but you will experience aliasing if your audio contains substantial energy above 30kHz (e.g. pulse and saw waves from analog VCOs). Instead, use the IN1-IN4 inputs to digitize audio that may alias above 30kHz."

that made me believe it would be better to use those ins…

Ah, so you were worried about aliasing (not lack of resolution) in your FM modulation. Is it useful to drive the modulation above 10-15kHz?

yes sorry, not my native language.
if its useful? for experimentaion :slight_smile: you never know until you discover something useful in the useless :blush: -
and… when you go in with a signal above our hearing range, and later you pitch the (in 96 or 192khz recorded) modulated sound down, interesting things can come to the light…

1 Like

And in fact an aliasing FM modulator might also yield interesting effects :nerd_face: However, I understand the desire for a pure and completely controlled setup even when experimenting. It enhances understanding and reproducibility.

By the way, you might get more use out of the IN1-IN4 inputs if you place a Fixed HPF afterwards, in order to remove the DC offset. Using the HPF (high pass filter) is equivalent to the AC-coupling that you will often see on linear FM inputs in the analog domain.


fantastic, thanks! thats a good tip.

(and yes, aliasing can also have interesting results, absolutely!)

Reviving this thread as I’m curious if it is normal that all 1-3 of my A, B, C and D inputs read 0c, -1c and -2c and some fluctuate between these values depending on the input, this is with nothing patched and simply just assigned to the V/oct of an oscillator.

never happened to me. no fluctuation of parameters if there’s no external stuff plugged in.

1 Like